Wednesday, June 06, 2007

TAMRON AF 17-50mm f/2.8 - Sharpness Evaluation

Ever since I have acquired the TAMRON SP AF 17-50mm f/2.8 XR LD ASPHERICAL [IF] standard zoom lens, it has been a 'kit' or 'default' lens that is on my earlier D50 and now, D200. It sure has given me plenty of photographs and helped me in the improvement progress. I do not claim that I am an expert now, but I definitely felt an improvement in my own skills and techniques.

Back then, when I was just using the D50, everything was good. But when I recently upgraded to the D200, that is when I start to have complaints. D200 being a whopping 10-megapixels, as compared to the D50 which is 6-megapixels, have a tendency to amplify any flaws of the lens that were hardly visible at lower resolution. I was pixel-peeping at my recent pictures and have realised that the in-focus subjects appeared rather soft. I had these problems during indoor shots, such as the
Old French Embassy and Istana Woodneuk series. These problems were also persistent even with flash photography during Vesak 2551 float parade. The recent indoor shoot at MoST also gave me plenty of soft pictures. Plenty of photoshop need to be done to overcome these soft pictures. I started to doubt the sharpness of my Tamron lens and believe that I had choosen a soft copy. Hence, I was thinking of selling away this lens, which I already have people waiting to view, and get myself the Nikon AF-S DX VR 18-200mm f/3.5-5.6 IF-ED*. I did not have to top up too much to upgrade from my existing lens to the VR lens, if I get a second-hand, which I also found. However, I have read countless reviews commenting on the sharpness of that lens**. Hence, in order not to make a mistake selling for a 'softer' lens, just for the sake of a longer reach and VR, I decided to give my Tamron lens a chance, a chance to perform. I did a simple test just by shooting some simple things in my room, to evaluate the sharpness of the lens.

(D200, Manual exposure mode, shutter speed 1/60s, L FINE JPEG quality, Customs setting default)
ISO has been adjusted accordingly when a different aperture was used.
All images are cropped at 100% without adjustment in sharpness. Levels were adjusted.

Figure 1. AF area is center. Focal point is neck area.
50mm at f/11

Figure 2. AF area is center. Focal point is the tetra-pack icon.
(a) Left, 50mm at f/2.8.
(b) Right, 17mm at f/2.8.

Figure 3. AF area is center. Focal point is the centre block of words.
17mm at f/2.8.

Figure 4. (a) Left, 17mm at f/2.8. Focal point is the tetra-pack icon. AF area is middle top.
(b) Middle, 17mm at f/2.8. Focal point is the middle block of words. AF area is center.
(c) Right, 17mm at f/2.8. Focal point is the block of Chinese characters. AF area is middle bottom.

The Verdict

From Figure 1, we can see that even the fine dust particles are visible and relatively sharp on Megatron. The face, neck, chest and Decepticon icon appeared very sharp. That is expected at f/11, which is thought to be the sharpess aperture for this lens and gives plenty of depth.

In Figure 2, the sharpness at the focal point is compared at (a) 50mm and (b) 17mm, since the aperture is fixed at f/2.8, giving a shallow depth of field. The focal point is the tetra-pack icon and in both pictures, the icon and the Chinese characters appear sharp.

In Figure 3, I shot a card with plenty of words at 17mm f/2.8. f2.8 was used since the card is flat and so all words should fall within a shallow depth of field. I engaged the shot at an angle, so there is some depth difference after all. However, on the whole, the focal point, which is the center block of words, appear sharp.

In Figure 4, I compared the sharpness at f/2.8, at three different auto-focus (AF) areas in the middle. (a) The top AF area is focused on the tetra-pack icon, (b) The middle center AF area is on the middle block of Malay and French words while (c) the middle bottom is on the lower block of Chinese characters. The sharpness is there for all the three AF areas.

My copy of the Tamron 17-50mm f/2.8 is definitely sharp and accurate without any front/back-focusing issues, and there is no doubt about it. So why are my earlier shots with the lens on my D200, plagued with soft images? I believe I have the right answers now. First of all, I must reiterate the 'Golden Rule of the Thumb to a Steady Shot'; when shooting without a tripod, in order to eliminiate soft/blurry pictures due to hand-shakes, shoot at a shutter speed that is '1/focal length'. For example, when I am focusing at 25mm, to get a steady shot, I must shoot at a shutter speed of 1/25 or faster.

During my shoot at the Old French Embassy and Istana Woodneuk, the lighting was low indoors. For the shots with flash and tripod, the images were sharp. But for the shots that are handheld, when engaging obscure angles, assuming that lighting was sufficient, the shots turned out soft. Those soft pictures were mainly shot at manual exposure setting with shutter speeds at 1/25 and 1/30s. I discovered that I can get rather steady shots at these speeds and not at any below 1/20s (based on individual). But I have forgotten that the 'golden rule of the thumb' here, hence any compositions that were at 35mm and longer, tend to be soft due to hand-shakes.

During the recent indoor shoot at MoST, in order not to risk tarnishing the toys with the strong flash light, I had respected the owner's request of not using flash. I was battling the noise at ISO800 even at the fastest aperture of f/2.8, while compromising plenty of depth. I did not have a tripod and hence, I was shooting everything handheld. Once again, I have forgotten the 'golden rule of the thumb'. I was trying hard not to exceed ISO800 to keep the noise minimum. Hence, I was trying various techniques to steady my hands in order to shoot at shutter speeds as low as 1/10s in AP exposure mode. Even when times when I switched to manual exposure mode to shoot at 1/25s (my comfortable shutter speed), I have forgotten that the 'golden rule of the thumb' was not applied, and was shooting at all focal lengths (even at 50mm). Hence, all these ignorance for the 'golden rule of the thumb', resulted in soft pictures.

What about flash photography? When using the flash, the flash is known to 'freeze' the action. By default, the minimum sync shutter speed in auto exposure modes (P, AP and SP) is set at 1/60s. At that shutter speed, the 'golden rule of the thumb' is applied to all focal lengths of my Tamron 17-55mm lens. However, during the Vesak 2551 float parade, I still get slightly blurry pictures even when I shot using AP exposure mode with flash sync shutter speed at 1/60s. I conclude that whatever that plagued my pictures, is the cause of Motion Blur. In order to 'freeze' action, 1/60 is good enough for slow moving subjects such as a slow-moving vehicle or walking. However, when it comes to fast-moving subjects, such as a highly-energetic dance, things start to get blurry. In order to 'freeze' such fast-moving subjects and eliminate motion blur, a shutter speed of at least 1/250s is necessary, which is also the maximum sync shutter speed of my SB-600 . I had not used such a fast shutter speed during the parade, as I would be sacrificing exposure, since the subjects were rather far and challenging the reach of my flash. To circumvent that, I would have to pump up the ISO and compromise on noise.

The Conclusion

My copy of TAMRON SP AF 17-50mm f/2.8 XR LD ASPHERICAL [IF] is without a doubt a sharp copy and no doubt about it. The problems of soft images is definitely a user problem. I have forgotten to apply the 'golden rule of the thumb'. I will withdraw the idea of selling my Tamron lens and pass on the Nikon 18-200mm VR lens at the moment. Meanwhile, I have learnt a good lesson today. With my D200, being a more advanced model, it requires the photographer to understand the principles well and master good techniques. I will have to work on that in order to produce better quality pictures.

* The 18-200mm is a decent 'walk-about' lens with 11.1x zoom that covers the range of my Tamron and a longer reach. With the Vibration Reduction (VRII) feature, the user can shoot at a shutter speed 4 stops faster. VRII works NOT by changing your shutter speed, but by stabilizing the image at shutter speeds that are normally subjected to handshake blur. Shutter speed affects exposure, but VRII will NOT affect exposure at all. One will not see a change in shutter speed during auto exposure mode such as AP and P, when VRII is active or inactive (do not quote me on this). One will only see a difference in the stable and sharp picture that is taken with VRII active. A trial with my Benny's copy (D50 AP, 200mm f/5.6 ISO400), demonstrated that it can get a sharp image compared to my Tamron (D50 AP, 50mm f/2.8 ISO400). At the above setting, my image with the Tamron was completely blurred by hand-shake, while the VR lens with VR active gave a sharp image at 1/15s steadily, however, even more blurry than the Tamron's if without VR (as expected due to the smaller maximum aperture and longer focal length). Although the VR lens has a minimum focus distance of 50cm, compared to Tamron's 27cm, the VR lens gave a larger reproduction or close-up of the subject.

** Most sites reviewed that the 18-200mm VR lens is soft. One trial between the 18-200mm VR lens and the pricey Nikon AF-S DX Zoom-Nikkor 17-55mm f/2.8G IF-ED, reviewed that at the same focal lengths and similar aperture settings (not the same as the lens constructions are different), the VR lens is significantly softer than the latter. This is as expected due to the fact that everything that gives the longer reach, all crammed into a lens that is smaller, sounds incredible! Tom Hogan gave a good review of both lenses (18-200mm VR, 17-55mm f/2.8). There are a few reasons for photographers toprefer having the f/2.8 than the 18-200mm VR (with maximum aperture at f/3.5 and f/5.6). The f/2.8 gives better and more natural background blur than f/3.5. f/2.8 also allows the lens to focus better in low light. f/2.8 allows a faster shutter speed than f/3.5, which makes a difference when shooting to 'freeze' moving subjects and eliminating motion blur, something that VRII cannot do. The variable aperture of the 18-200mm VR lens during zooming, creates an inconvenience in shooting, since aperture affects exposure as well, apart from depth of field. The change in aperture will result in a change in exposure, hence the other settings such as shutter speed or ISO will have to be adjusted accordingly.


Constructive comments and suggestions are welcome.
Thank you.

No comments: